1

From whispering to shouting

A taxonomy of RRPs’ blurring strategies on economic policy after 1990

Matteo Cavallaro

Centre d’Economie de l’Université Paris Nord (CEPN)

Abstract

Our research aims to provide a global analysis of the differences between European RRPs’ positions

on economic matters with a specific focus to the attention paid by the different RRPs to these

issues. We do so by analysing data from the MARPOR project database that has been treated

according to Lowe et al.’s (2011) transformation. We cover 42 different RRPs across 22 European

countries over a period ranging from 1990 to 2017. Our findings show that economic issues are still

less important in RRPs programmes than in those of other political families. However, the gap has

been reducing as RRPs, in their programmes, talk more about the economy than before.

The article then describes RRPs’ heterogeneity on economic issues. We draw a cartography of the

‘economic policy space’ within the radical right party family. We do so by the means of various

statistical methods (PCF and Ward’s method) applied to manifestos data from the MARPOR project

database. We find four dimensions that can describe large part of the differences between RRPs on

economic matters. These dimensions all describe the “salience” of a set issues and we labelled them

‘Welfare’, ‘Economic Liberalism’, ‘Economic Management’ and ‘Protectionism”’ Our results show

that RRPs are still very different in their economic programmes in terms of contents and, using

Ward’s clustering method, we regroup RRPs in four classes: ‘neo-liberal’, ‘pro-welfare’, ‘blurring-

by-muting’, and ‘blurring-by-multiplying’. In particular, we highlight a trend from a ‘blurring-by-

muting’ towards a ‘blurring-by-multiplying’ strategy by most RRPs on economic matters. We

conclude with a preliminary account of the possible determinants of this change of strategy and find

that the economic situation does not seem to be a valid explanation, while the adoption of similar

strategies by non-RR parties appears to be the strongest predictor.

Paper prepared for the workshop on Radical Party Challenges and Realignment in Advanced

Democracies,

Duke University, April 2018

Early draft. Please do not cite or distribute without authors’ permission.

2

1 Introduction

How do Radical Right Parties (RRPs) in Europe differ on economic matters? This question has

been the subject of renewed debate in political science following the emergence of new and strong

RRPs in the 1980s, identified by von Beyme (1988) as a ‘the third wave’ of RRPs developments.

As of now, there seems to be a consensus that economic matters are secondary and instrumental

issues for RRPs, the parties being defined by an ideology blending nativism, authoritarianism and

populism (Mudde, 2007). Consequently, RRPs are considered to compete mostly on the non-

economic dimension, which explains the adoption by these parties of a blurring strategy on

economic matter, i.e. RRPs either avoid talking of economic issues or adopt ambiguous positions on

these matters (Rovny 2013). Indeed, it seems impossible to give a clear answer to the question

‘where do radical right parties stand?’ (Rovny 2013): the literature underlines that RRPs' strategy

on economic matters (anti-tax, pro-welfare, etc.) has been varying since the 1990s in order to satisfy

an electorate with heterogeneous economic positions (De Lange, 2007; Ivarsflaten, 2005): silence

and heterogeneity are therefore the main characteristics of RRPs economic programmes.

Nevertheless, we may assume that economic matters have become increasingly important in RRPs’

messages after the 2008 financial crisis. Indeed, the context is characterized by worldwide major

economic and social backlashes while the electoral weight of the RRPs have risen significantly in

many countries over the period (Hernández and Kriesi, 2016). Given the importance of the

economic crisis in the political system, we may hypothesise that economic issues are becoming

increasingly important also for RRPs. Similarly, the adoption of austerity policies by many

European countries might have opened the route for RRPs to become the defenders of redistribution

policy, coupling the defence of the public service with RR nativism through the adoption a ‘welfare

chauvinist’ appeal. In other words, RR might be abandoning the blurring strategy that allowed them

3

to gain a large success in the last years and there might a converging trend within the party family

towards a common stance.

In order to understand whether RRPs heterogeneity has been decreasing, this article reconsiders the

differences among RRPs to enable comparison across countries and time. In line with the literature

(De Lange, 2007; Rovny, 2013; Ivaldi, 2013), we rely on party manifestos as the source of RRPs'

programmatic positions. As a result, our work brings three types of contributions to the literature on

RRPs’ economic programmes. To our knowledge, it is the first work which proposes a taxonomy of

the differences between RRPs specifically on economic aspects and through both diachronic and

geographic dimensions. Second, this is done using recent data, thus taking into account the effects

of the crisis which started in 2007-2008 and in particular the importance assumed by economic

issues in RRP programmes following the great recession. Third, we propose an original

methodology to explore and analyse manifesto data, based on inductive analysis of transformed

data, following Lowe et al. (2011) and Prosser (2014).

The article is structured as follows: Section 2 provides a review of the literature on the topic

introduces our hypothesis.. It highlights the main results for RRPs' positioning on economic matters

and the limits of the literature with this respect. Section 3 discusses and justifies our choice to focus

on party manifestos. Section 4 presents the methodology and section 5 describes the dataset and

justifies the list of RRPs considered in our work. Section 6 presents the results and section 7

concludes summarising our contributions but also discussing the limitations of our work and the

possible agenda for further research.

2 RRPs “schizophrenic” agenda on economic issues: an evolving

strategy to compete in a multidimensional political space

Scholars seem to agree RRPs’ position on the secondary role that economic policy plays in RR

ideology (Mudde 2007) as they underline the strong variations that occurred in RRPs’ stances over

4

time. Indeed, if we look at the evolution of the literature, we can distinguish three major replies to

the question “Where do RRPs stand on economic policy?”.

Early studies (Kitschelt, 1995) pointed out that RRPs used neoliberalism and free market to attract

voters from the middle class, while nationalism was key to obtain the support blue collar workers.

This combination has been described as the RRPs' ‘winning formula’ and has been considered

among the reasons behind the emerging of a successful ‘third wave’ of RRPs and during 1990s

found support in different authors ((Betz, 1994; Ignazi, 2003).

In a following analysis, Kitschelt (2004, p.10), however, argued that, following changes in the

social composition of their own electorate, RRPs were toning down ‘their market liberal rhetoric so

as not to alienate potential working class constituencies’. De Lange (2007, p.429) empirically

addresses the new ‘winning formula’ proposed by Kitschelt (2004), and compares three different

RRPs – the French Front National (FN), the Flemish Vlaams Blok (VB) and the Dutch List Pim

Fortuyn (LPF) – and concluded that the FN and the VB held a fairly centrist position on economic

questions in the new millennium', while the Dutch LPF was still neo-liberal in its economic

proposal.

This shift goes even further than this ‘centrist’ appeal: Ivaldi (2015b, p.362) finds that the FN has

recently been moving towards a ‘new leftist-agenda’ that is ‘increasingly framed by economic

egalitarianism’ and ‘marked by the acceptance of state regulation and public services, and

emphasizes income redistribution’. This goes together with an increased importance paid by the

French RRP to economic issues (Ivaldi, 2015a). Thus, the FN is becoming close to what Kitschelt

(1997) identifies as welfare chauvinism, that is, the ‘combination of a left-wing position on the

economic axis and a authoritarian/exclusionist position on the cultural dimension’ (Ivaldi, 2013,

p.14). Welfare chauvinism is also part of the True Finns’ (PS) programme (Keskinen, 2016) and

characterizes the Dutch radical right electoral constituency (De Koster, Achterberg and Van der

Waal, 2013).

5

The existence of different economic stances within the RRP family eventually led to questioning the

actual importance of economic issues in RRPs’ political discourses. According to Mudde (2007,

p.119), ‘the economic program is a secondary feature in the ideologies of populist radical right

parties’. The many shifts in RRPs' economic programmes rest on the instrumental nature of

economic policy for these parties, which underlies their ‘schizophrenic socioeconomic agenda’

(Mudde, 2007, p.135). Developing on this approach Rovny's (2013, p.5) argues that, since RRPs

compete mainly on non-economic dimensions, these parties adopt a strategic blurring on economic

issues, that is the fact of taking ‘vague, contradictory or ambiguous positions’ on certain issues, in

order to attract a larger audience. In particular, Rovny (2013) distinguishes three main types of

‘position blurring’: (i) the muting of harmful issues (that we call ‘blurring-by-muting’), (ii) the

multiplicity of position (that we call ‘blurring-by-multiplying’), (iii) and an incoherent position over

time. His results show that RRPs have been adopting a blurring strategy, in particular as they have

been constantly underemphasizing economic issues in their economic programmes. In order words,

in order to attract the a larger audience unified on non-economic issues (e.g. immigration or law and

order) but divided on economic policy, RRPs have been avoiding to take a stance by being silent on

the topic. However, some RRPs appear to be assigning more importance to economic issues in the

most recent period and, specifically, following the subprime crisis. For example, Ivaldi (2014) finds

that the FN paid more attention to economic issues in the 2012 election and ‘has accentuated its

nationalist anti-globalization stance while endorsing left-wing economic policies’ (Ivaldi, 2014,

p.14). Indeed, following 2008 economic crisis, RRPs might now have an incentive in taking

position on economic policy, blaming unregulated markets as the culprit for financial instability,

globalisation as the source of country’s economic disarray, and mainstream parties for the adoption

of austerity. As such, taking a stronger stance on redistribution might be strategically convenient to

expand its electorate towards the so-called ‘losers of globalisation’, ‘those who are unable to cope

with the "acceleration of economic, social, and cultural modernization" and/or are stuck in full or

partial unemployment, run the risk of falling into the new underclass and of becoming "superfluous

6

and useless for society"’(Rydgren, 2007, p.248). This leads us to our first hypothesis. Following

Rovny (2013), we assume that economic issues are less relevant for RRPs than for other parties.

However, we argue that there has been an absolute and a relative increase trend in their importance,

as Ivaldi (2014) already shown for the French FN:

H1: The salience of economic issues is overall lower for RRPs than for other parties, but

with a reducing gap.

Finding support for this prediction would legitimate our research question insofar as this would

mean that economic matters have become strategically more important for RRPs.

As H1 focuses on the salience of economic issues in general, we are left with no clue on which

issues now receive more emphasis in RR programmes. Indeed, the divergences within RR social

basis on economic issues that led to the blurring of their position on economic policy are still

present: as Bornschier and Kriesi (2012) underlines RR electorate is still composed of both small

business owners and production workers, social groups that appear to be divided on economic

matters. On the other hand, Oesch and Rennwald (2017) underlined that despite the strong

penetration of RRPs among small business owners, these account for only 12% of the RR

electorate. The smaller numbers of the ‘pro-market’ part of their constituency and the openings

provided by 2008 financial crisis in penetrating among production (and service) workers, the RRPs

might work as an incentive to abandon a blurring strategy, to have a clearer stance on economic

policy. As such, in our Hypothesis 2 we consider whether RRPs abandoned position blurring or

simply moved from a blurring-by-muting to a blurring-by-multiplying strategy:

H2: The increase in salience of economic issues is accompanied by a decrease in the

position blurring of RRPs.

Once ascertained the main characteristics of the strategy of RR party family on economic policy in

terms of salience and position blurring, we move to describe the diachronic and cross-country

heterogeneity of RRPs’ strategies. We do so as we are interested in analysing the differences

between RRPs in terms of blurring strategy. Indeed, position blurring might be pursued in different

7

ways, for example by paying more attention on certain issues, which are probably perceived as less

problematic than others. An example of such a change is the FN, as identified by Ivaldi (2015), as

the French RRP seems to still keep both pro-market and pro-welfare issues in its programme,

despite shifting the attention from the former to the latter. We argue that a similar change can be

seen in the analysis of party manifestos:

H3: Within the context of a position blurring strategy, the composition of RRPs’ manifestos

on economic matters have significantly changed over time, as pro-market elements receive a

decreasing attention to focus more on welfare issues.

Finally, we turn our attention towards the determinants of RRPs strategy on economic matters and

propose a first assessment. The literature underlines that the choice of emphasising certain issues

also depends on particular characteristics of the political system or other contextual factors (XXX).

In particular, Green-Pedersen and Mortensen (2010) argue that parties need to respond to issues

raised by other political forces, the so-called ‘party system agenda’. Even if competing on a non-

economic dimension, RRPs might therefore be ‘forced’ to either increase the salience of economic

matters because this is at the centre of the party system agenda. Similarly, we can imagine that

RRPs position blurring is influenced by other parties blurring strategies: if other political forces

avoid taking a clear stance on these matters, RRPs will probably have an incentive in blurring even

more their position. As such our hypotheses 5a and 5d are as follows:

5a: We expect RRPs to pay more attention to economic matters in elections where these issues

are more important for other political forces.

5d: We expect RRPs to increase the ambiguity of their position on economic matters in

elections where other political parties adopt blurring strategies as well.

Spoon et al. (xxx) underline that a party’s success is also a possible determinant of salience strategy.

In line with this idea, Akkerman et al. (2016) analysed whether RRPs were ‘mainstreaming’, that is

whether they were changing their profile following the electoral success of the third wave.

Similarly, we consider that party size could be a valid predictor of both economic salience and

8

position blurring. Larger parties are supposedly less keen to leave aside relevant dimensions of

issue competition (Spoon, Hobolt and Vries, 2014): as such, we expect more successful RRPs to

pay more attention to economic issues. On the opposite, however, we imagine that the electoral

success of RRPs should lead them to adopt a less clear position on economic issues. In other words,

larger RRPs should be moving from a blurring-by-muting to a blurring-by-multiplying strategy, as

they cannot avoid the issue, but can still avoid taking a clear stance. In terms of what makes a party

successful, we should consider two options (i) party’s electoral score and (ii) party’s incumbency,

which, as underlined by Vavreck (2009) has an effect on the salience of economic issues:

5b: We expect more successful RRPs to pay more attention to economic matters.

5e: We expect more successful RRPs to increase the ambiguity of their position on economic

matters in elections.

Finally, the degree to which the incentives a party has in emphasizing the economy are not

determined solely by the characteristics of the party system, as political actors also react to the

broader economic context. As Spoon et al. (2014, p. 368) ‘during times of economic hardship,

economic issues are likely to dominate the political debate, and parties will emphasise these issues

in their election manifestos.’ As such, we expect RRPs pay more attention to economic issues when

the economy slows down. However, this is not necessarily true for position blurring: as economic

crisis might have an impact on different parts of the society, RRPs still have a clear incentive in

blurring their position to attract a large audience.

5c: We expect RRPs to pay more attention to economic matters in elections where these issues

are more important for other political forces.

5f: We expect RRPs to increase the ambiguity of their position on economic matters in elections

where the economy experiences a slowdown.

9

3 Data and methodology

3.1 A supply-side approach based on a manifesto analysis

The literature identifies two main approaches to analyse party differences: a demand-side approach,

which studies parties’ positions according to voters’ preferences, and a supply-side approach, which

focuses on what parties offer voters, either ideologically or in terms of programmes (Kriesi et al.,

2008; Benoit and Laver, 2006). Since our aim is to study differences in the economic programmes

of parties in the RRP family, we favour a supply-side approach. We can distinguish between three

subgroups of supply-side approaches:

- Expert surveys, which rely on scholars’ judgements (Castles and Mair, 1984);

- Analysis of party’s parliamentary activity, in particular voting behaviours (Hix, Noury

and Roland, 2007);

- Manifesto analysis, in particular through the MARPOR project (Volkens et al., 2015) and

the Euromanifesto Study (Braun et al., 2016), which provide human-coded quantitative

data on party manifestos.

Expert surveys are widely used to analyse differences among parties. However, surveys introduce

the non-negligible risk of reproducing the general views in the literature on the topic. Experts do not

necessarily have first-hand knowledge on every party and mostly provide confirmation of the main

literature. As Kriesi et al. (2008) argue it is impossible to know how experts form their own

opinions: in other words, rather than a taxonomy of parties, they provide a taxonomy of the

‘received wisdom’ (Mudde, 2007, p.295).

Moreover, Kriesi et al. (2008) underline the difficulties facing experts who study secondary issues.

This applies to the case of RRPs’ positioning on economic matters, which is not a central tenant of

their ideology.

More rigorous analysis depends on an analysis of coded manifestos and parliamentary activities (i.e.

voting records and parliamentary speeches). In this article, we rely on coded party manifestos as

10

various authors in the literature. Indeed, Norris (2005) studies the consequences of RRPs on

national political systems, Arzheimer (2009) and Arzheimer and Carter (2006) analyse the impact

of the positions of other political competitors on voters' support for RRPs. Cole (2005) assesses the

ideological differences between ‘old’ and ‘new’ RRPs. Also the already cited works by De Lange

(2007), Rovny (2013) and Ivaldi (2014) rely similarly either on a specific existing dataset or on the

authors' self-coding. This type of approach allows us to consider RRPs’ economic positions on the

period ranging from 1991 to 2017, for all European countries including Eastern European countries.

3.2 Presentation of the MARPOR database

In this study, we use the dataset provided by the Manifesto Research on Political Representation

(MARPOR), formerly the Manifesto Research Group and Comparative Manifesto Project (Volkens

et al., 2015). This dataset provides information on the content of almost 4,000 political manifestos

presented at national elections, covering 988 parties in 56 countries, not exclusively European. Data

are hand-coded by dividing each manifesto into ‘quasi sentences’ or non-overlapping textual units

which refer to a policy proposition. These units are classified into seven domains, each of which is

split into policy categories (and eventually subcategories) for a total of 56 mutually exclusive items.

The dataset reports the relative number of quasi sentences for each item in each manifesto, that is,

count of mentions of a certain category divided by the total count of quasi-sentences.

3.3 The MARPOR database and salience issue

The coding of the dataset follows a particular approach according to which parties do not compete

on every issue, rather they ‘differentiate themselves by emphasizing the issues on which their

stances are most credible’ (Lowe et al., 2011, p.133); this emphasis is referred to as salience or

valence. The MARPOR dataset counts the mentions of a certain issue, thus, providing a measure of

the importance (salience) of the focal issue to a certain party.

11

This is in line with Budge's (2001, p.82) idea that differences among parties ‘consist of contrasting

emphases placed in different policy areas’ and, in particular, that all parties sustain a certain set of

undisputed issues. For example, ‘MARPOR item 501 – Environment’ provides us with information

on a party manifestos' number of positive references to environmental policy, but does not include

information on negative references. The assumption is that parties favouring environmental

protection will refer frequently to it, while parties that oppose stricter environmental laws will

probably not mention it.

The opposite approach, referred to sometimes as ‘direct confrontation’ (Dolezal et al., 2014),

follows Downs's (1957) conceptualization of the political space. According to him, parties compete

by presenting alternative policy options for each major topic. In our example, parties would

compete on environmental policy by presenting clear alternative options, rather than emphasizing or

de-emphasizing the issue. Other authors have also criticised the salience theory. For instance, Laver

(2001) underlines that, in the phrase ‘ If we win the next election all bankers will be taken out and

shot’, what is relevant is the underlying positional idea, not its repetition throughout the manifesto.

Franzmanna and Kaiser (2006) show that, a political system includes both positional and salience

issues. Also, political systems change over time and the type of competition over the same issue

might be considered confrontational in an election and salient in another. Dolezal et al. (2014)

tested various salience theory assumptions using data from the 2002, 2006, 2008 Austrian national

elections. They found that most do not hold. However, they also found support for the thesis that

parties that ‘own’ an issue (such as Green parties and environmental policy) emphasize the topic in

their programmes.

Despite the shortcomings highlighted in the literature in relation to salience-based data, the use of

MARPOR is appropriate in our case for at least two reasons. First, the dataset is not strictly

‘salience-based’. McDonald and Mendes (2001) pointed out that MARPOR codes a large array of

categories in a confrontational way, providing both positive and negative items (e.g., item 406 –

‘Protectionism: positive’ and item 407 – ‘Protectionism: negative’). Also, salience and

12

confrontational approaches are not mutually exclusive inasmuch as they stress different aspects of

political competition (Budge, 2001). Thus, it is possible, through a simple operationalization, to

reduce the ‘salience approach’ in MARPOR data in order to take account also of the confrontational

dimension (Gemenis, 2013). Second, our research question deals in part with the salience of

economic matters in RRPs’ programmes: before moving to a party position, we need to address the

importance of economic issues for those parties. Salience is especially important for the present

work, while the confrontational scale might be less relevant for RRPs in the context of economic

issues. Moreover, the confrontations among positions is more difficult to handle since RRPs are

characterized by ‘schizophrenic agendas’, which can include both free-market and state intervention

positions together. As such, the aggregation of pro-state and pro-market pledges (which can each be

extreme) would result in a centrist stance, hiding the contemporary presence of opposite stances.

This is exemplified by FN’s economic position on Benoit and Laver’s (2007) socio-economic ratio

scale which appears to be on the centre-right1, thus losing the information on the possible

simultaneous presence of ‘left’ and ‘right’ elements in its programme.

For all these reasons, we compare the salience given to economic matters by each political family,

relying on the MARPOR dataset. We also consider specific operationalisations of the data in

relation to our question.

4 Methodology

The operationalization of MARPOR data has been the subject of major debate since the data

transformation is aimed at producing one or multiple scales to describe the dimensionality of the

political space. The literature proposes two main methodologies: inductive and deductive

techniques (Prosser, 2014).

1 In Benoit and Laver (2007), the socio-economic position of the FN in 1993 and 1997 is, respectively, .32 and .17 on a

scale from -1 (left) to +1 (right).

13

Deductive techniques are based on a priori selection of components from which one or many

indexes can be deduced using ‘common sense substantive judgements to sort the component

variables of the scale’ (Benoit and Laver, 2007, p.100).The best known example of a deductive

technique is the RILE scale (Budge, 1999), an unidimensional description of the policy space based

on the difference between the sum of left-wing and right-wing quasi-sentences found in the

MARPOR dataset, divided by the total number of quasi-sentences. Benoit and Laver (2007)

developed a bi-dimensional scale to describe the left/right differentiation on both social policy and

economic dimensions. Lowe et al. (2011, p.131) proposed a ‘logit scale of position’, to restructure

the dataset as if each additional reference to an issue had a decreasing marginal impact, before

considering left/right differentiations. This decreasing marginal impact is supported by the Weber-

Fechner law of psychophysics which suggests that ‘we should operate in proportions, not levels,

and work with a logarithmic scale relationship between the underlying quantity and subjective

estimations of it’ (Lowe et al., 2011, p.130). However, as argued in the previous Section, deductive

techniques rely on the author’s judgment about the left/right categorization of each item of the

database and might be too normative to understand the differences between RRPs.

Inductive techniques seek to identify latent dimensions and make fewer assumptions about how

they are related. Such techniques rely on exploratory statistics such as Factor Analysis (Alonso and

Volkens, 2012, pp.78–80) and are useful if there is some doubt about the relationship between the

variables, as in our case.

The literature warns about the limits of inductive techniques. Prosser (2014) points out that these

methods applied to manifesto data usually find a rather high number of factors/components with

low eigenvalues. When some authors decide to limit their choice to the first and sometime second

component (Petry and Pennings, 2006; Gabel and Huber, 2000), the part of variance explained by

the retained factors/components is often very low. According to Gemenis (2013) and Prosser

(2014), this problem is highlighted particularly by post-estimation tests such as the Kaiser-Meyer-

Olkin (KMO) index and Bartlett’s test of sphericity.

14

Limitations related to sampling adequacy can be overcome by employing some easy data

transformations. In particular, logarithmic transformation is acknowledged to be a good pre-

treatment for data before conducting factor analysis (Wold, Esbensen and Geladi, 1987; Aitchison,

1986). Moreover, this transformation is in line with Lowe et al. (2011), where each additional

reference to a topic in the manifesto has decreasing marginal importance for the reader (and the

authors).Since it has been proven to improve the validity of manifesto data (Kim and Fording,

1998), we retain this assumption when analysing the data. However, as underlined by Prosser

(2014, p.96), a simple logarithmic transformation has some relevant shortcomings, in particular “not

mentioning an issue at all in a small manifesto may score higher than mentioning an issue in several

sentences in a large manifesto”. While this is not a problem when comparing issues in a single

manifesto (e.g. the issues cited more often will have a higher score, thus keeping the order of

importance), it becomes an element of distortion in comparing different manifestos. Therefore,

although we agree with the theory in Lowe et al. (2011), we use a modified version of the log scale,

as proposed by Prosser (2014), which improves the comparability of different manifestos:

𝜃(𝐼) =log(𝑆 + 1)

log(𝑁 + 1)

where S is the number of quasi-sentences on a given issue I and N is the total number of quasi-

sentences in the manifesto. The index ranges from 0, for party manifestos that have no references to

economic matters, to 1 when the entire programme is dedicated to economic issues.

Our analysis then proceeds as follows. First, in order to test Hypothesis H1, we analyse the

evolution of the log scale of salience presented above. We obtain two types of scales: (i) a scale of

economic salience by aggregating all economic items and (ii) a specific scale of salience for each

separate economic item.

Second, in order to test Hypothesis H2, we develop an index of economic issues heterogeneity

based on Benoit and Laver (2007) socio-economic scale. We consider the contemporary presence of

‘left-wing’ and ‘right-wing’ items as a good indicator of position blurring: if a party pays the same

15

attention to pro-market and interventionist elements it does so either because it is taking a centrist

stance, or in order to avoid taking a clear stance on an issue. The index if formalized as follows:

𝐻𝑒𝑡 = 𝑀𝐼𝑁(𝜃(𝐿); 𝜃(𝑅))

𝑀𝐴𝑋(𝜃(𝐿); 𝜃(𝑅))

Where 𝜃(𝐿) is the salience of left-wing items and 𝜃(𝑅) is that of right-wing items as categorised by

Benoit and Laver (2007) and calculated relying on the previously presented log scale as proposed

by Prosser (2014). The index ranges from 0, for parties including only items pertaining to left/right-

wing economic policy, to 1 for parties whose programme present the exact same amount of space

for left-wing and right-wing pledges on economic matters.

Then, in order to check the validity of Hypotheses H3 we run a principal component factor analysis

(PCF) of party programmes (considered as individuals) from all election years, the variables being

the economic items and the log scale of salience as the values within each cell in the table. This

allows us to project the position of a party’s programme in different years in order to compare the

evolution of party’s proposals through time. To ease interpretation, we apply normal Varimax

rotation (Kaiser, 1958), a common criterion for orthogonal rotation. Finally, we regroup RRPs’

programmes using Ward’s clustering method (1963) on the same data. This method allows us

obtain a typology minimising the variance within each class and maximising the variance between

the classes. We interpret these groups using the average position of each class on the factorial space.

Finally, in order to test our hypotheses H5a – H5f, following the example of Spoon et al (2014) we

use a Prais-Winsten transformation technique for OLS with robust standard errors as this method

corrects for auto-correlation problems in panel data avoiding the shortcomings of adding a lagged

dependent variable. To account for potential differences across countries, we also include country

fixed effects in all models. We distinguish between two dependent variables. In models H5a-H5c,

as we wish to explain the level of economic salience we rely on the previously described modified

version of the log scale, while in models H5d-H5f the dependent variables is the index of issue

heterogeneity. In order to test the impact of RRPs’ incumbency, we create a dummy that assumes

16

the value of 0 for RRPs contesting the election as opposition forces, and 1 if RRPs have been part of

the incumbent cabinet or have offered external support. The coding is based on Akkerman (2012),

Minkenberg (2015) and own research. Concerning electoral scores, data are obtained directly from

the MARPOR dataset (item ‘pervote’). Finally, to test the influence of the economy on the salience

of economic matters in RR manifestos and on position blurring, we rely on economic growth as

calculated by Eurostat, which measures the year-to-year real GDP growth rate.

5 Data selection and description

Since our objective is not to analyse the competition among all the political forces, but rather to

identify areas of RRPs disagreement, and its importance. Thus, we focus only RRPs and exclude

other parties.

The definition and identification of RRPs is not straightforward and there is a large literature on this

issue. We adapt Mudde's (2007) definition and party selection. Some of these adaptations refer to

updating the 2007 list relying on Mudde (2014) and thus including the Dutch PVV, the Greek

Golden Dawn (XA), the Swedish Democrats, the Hungarian Jobbik and the Latvian National

Alliance. We also considered two borderline cases which we include in RRPs: the True Finns (PS)

and the Dutch List Pim Fortuyn (LPF). Indeed, the more recent literature includes the Finnish

movement in the RRP family. Arter (2012, p.815) notes that, ‘the ethno-nationalist ideological

dimension has become more pronounced in the PS’s policy output’ and the PS has adopted a

welfare chauvinist economic programme (Arter, 2010, p.499). Raunio (2012, p.20) also concludes

that ‘the empirical analysis of The Finns’ EU policy in turn has shown that their discourse is overall

broadly similar with other European populist or radical right parties’. The radicalism of the no

longer existing List Pim Fortuyn is more debated. Mudde (2007, p.47) categorizes it as ‘non-radical

right populist’. However, according to Akkerman (2005, p.350), on the one hand, ‘the LPF program

has a substantial overlap of liberal and nationalist principles’, and on the other it adheres to a form

of civic nationalism not so distant from the other parties in our list: the LPF's strong anti-Islam

17

position was a core feature of party’s ideology, as acknowledged by Akkerman (2005, p.341).

However, apart from Islamophobia, the LPF's remaining ideology is hardly ‘radical’ and favours

assimilation (rather than integration) of foreigners (Rydgren, 2004b, p.17).

The list of RRPs is provided in Appendix 1. It includes 42 RRPs across 22 European countries over

the period 1990 to 2017. This gives us a total of 122 party manifestos.

Among economic items, MARPOR includes two domains, which are clearly related to our work:

domain 4 - ‘Economy’ and domain 5 - ‘Quality of life’. However, within these domains, not all

items refer to policy stances. Some are generic and unspecified economy-related items such as item

408 - ‘Economic Goals’. Others cover both economic and non-economic issues and should thus be

excluded in order to avoid misleading data. This is the case for item 411 - ‘Technology and

infrastructure’ and item 501 - ‘Environment’. Conversely, we should not exclude item 701 –

‘Labour groups: positive’ (positive judgments about trade unions) even though it is coded under a

non-economic domain.

We rely on Laver and Benoit (2007) to identify the set of variables for the economic dimension, but

without making any assumptions about their relation to one another. On this basis, we build on their

economic policy ratio scale to generate a new variable as the logarithm of the sum of all quasi-

sentences of selected items divided by the logarithm of the total number of quasi sentences of the

manifesto. We therefore obtain an index of economic salience, or the importance of economic

proposals in the party programme.

Table 1 summarizes the information on each variable in domains 4 and 5 for RRPs. This allows us

to identify the most important topics. The highest value for the policy importance index is for item

504 ‘Welfare State Expansion’, while neoliberal policies such as item 401 ‘Free Market’ and item

402 ‘Incentive’ are less frequent. There are more positive than negative references to protectionism,

while the least important item is 415 – ‘Marxism’, which figures in only 3 of the 122 programmes

analysed. A very limited place is left to item 413 – ‘Nationalisation’, however, items 412 –

18

‘Controlled Economy’ and 405 – ‘Corporatism’ appear rather close, thus this result might be a

consequence of coding decisions.

*** Table 1 AROUND HERE ***

6 Findings

6.1 The salience of economic policy for RRPs: an increasing trend which does

not modify RRPs’ blurring strategies

Table 2 presents a comparison of our index of economic salience across different party families.

Our results indicate that RRPs pay less attention to economic policy than do the other main political

families, while the standard deviation is higher (suggesting that RRPs differ in the importance they

assign to the economy).

*** Table 2 AROUND HERE ***

Figure 1 depicts the historical evolution of economic salience for RRPs compared to liberal,

conservative, communist and social democratic parties. RRPs show the strongest increase, starting

from 0.58 in 1990-1994, to 0.74 in 2010-2017. Although the level is lower than that of other major

political families, the difference is becoming less evident. This is also due to the fact that, in the

1990s, some parties (e.g., the Italian MSI in 1992, and the Croatian HDZ in 1995) made no

reference to the economy while, after 2000, all RRPs include economic items in their programmes,

with an increasing number of included issues. Figure 2 depicts the closing gap with other political

families: in the early 1990s RRPs included on average fewer than 6 economic items, rising to 8.4 in

the 2000s, in line with the communist and liberal parties2. It indicates a clear re-alignment of RRPs

with the other political families for the importance assigned to economic matters through time, thus

supporting Hypothesis H1. However, this realignment does play out differently in Western and

2 This result holds also if we consider all the items included in domains 4 and 5 of the MARPOR dataset. Here, the

average number of items increased from fewer than 9, to 12, the largest increase for a party family.

19

Eastern Europe, as shown in Figure 3 and Figure 4. While we do see an increase in the salience of

economic issues in both sides of Europe, in Western Europe the trend is weaker and the difference

between party families appears clearer.

*** FIGURE 1 TO 4 AROUND HERE ***

The increasing salience of economic matters in RR party manifestos is not, in se, sufficient to

conclude that RRPs now take a stance on economic matters. In order to assess whether RRPs have

modified their position blurring on economic matters, we look at the evolution of our index of

economic issues heterogeneity (Figure 5). Despite a decrease in the 2005-2009 period, which

depended exclusively on two parties, the Danish DF and the Bulgarian ATAKA having part of their

manifestos entirely dedicated to right-wing/left-wing economic policy, RRPs appear to still include

policies from both sides of the spectrum. As discussed in section 2, position blurring can appear in

different forms, either as a lack of a position or as the contemporary presence of positions of a large

number of positions. It would seem that, despite the increasing attention that RRPs appear to pay to

economic issues, the clarity on their positions, at least measured using a left-right economic scale,

has not increased. This contradicts our H2 and indicates that RRPs have not abandoned their

strategy of position blurring on economic matters.

** FIGURE 5 AROUND HERE **

Finally, the data show that the difference between RRPs in terms of salience of economic items has

also been decreasing: we tested the decrease in standard deviation by applying Levene’s test; it was

significant at the 0.05 level. This appears to be preliminary evidence of H3, namely the existence of

a possible intra-family converging trend on a common strategy when it comes to the importance of

economic issues. This is addressed in the following section by exploring the dimensionality of

RRPs' economic policy.

20

6.2 The dimensionality of RR economic policy

In order to analyse the inherent dimensions of RRPs' economic policy supply, we perform a PCF on

all RRPs based on the emphasis they give to the issues identified by Laver and Benoit (2007).3 As

preliminary validation, we test the transformed dataset with KMO and Bartlett's test of sphericity.

An important consequence of logarithmic transformation is that the KMO measure of sampling

adequacy increases to 0.8, a value well above the minimum threshold.4 Having ascertained that our

dataset is fit for a PCF, we present the results.

6.2.1 The main factorial components structuring RRPs on economic matters

*** Table 3 AND Table 4, FIGURE 6 AROUND HERE ***

Table 3 reports the Eigenvalues and variance for each component. The first component describes

33% of the variance and, taken together, the four first dimensions explain 64% of the total variance.

Table 4 reports the loadings following Varimax rotation and

Figure 6 depicts the correlations between the variables and the components.

‘Free Market’ (401), ‘Welfare State Limitations’ (505), ‘Incentives’ (402) and ‘Economic

Orthodoxy’ (414) are positively correlated to the first dimension meaning that the second axis

describes the importance given by the party to economic liberalism. This takes the form of welfare

cuts or privatizations, but also tax breaks and policies favouring individual entrepreneurship.

The second component, instead, divides RRPs according to the salience given to items ‘504’

(‘Welfare State Expansion’), ‘506’ (‘Education expansion’) and ‘403’ (‘Market Regulation’).

Interestingly, items 505 and 401 (free market and welfare state limitations) do not load negatively

onto this dimension. This means that RRPs might display pro and anti-welfare positions

simultaneously. Thus, the second axis is an index of ‘welfare’ salience in the manifesto rather than a

3 Due to its extremely low number of appearances, we decided to drop item “406 – Protectionism : negative”. 4 Although convenient, this consequence should not be the main reason to perform such transformation. What really

matters is the underlying hypothesis, namely, that each supplementary sentence on a certain issue is “less important” for

evaluating the salience of that content for the party. Logarithmic transformations resolve one of the main limitations of

inductive approaches, however, this is still conditional on the appropriateness of the theoretical framework.

21

confrontational dimension, and it opposes RRPs that mention welfare in their programmes against

RRPs which do not.

The third component resembles what Singer (2013) calls the ‘economic management’ dimension.

This refers to the salience of policies advocating a state-controlled economy and is represented by

items 412 – ‘Controlled Economy’, 413 – ‘Nationalisation’ and 404 – ‘Economic Planning’. Again,

there is no clear opposition to a free market; it is more of a salience indicator than a confrontational

dimension. Finally, the fourth dimension is structured by RRPs’ position with respect to

protectionism (item 406 – ‘Protectionism: positive’).

*** Figure 7 around here***

We regroup RRPs manifestos into clusters using Ward’s (1963) method. Looking at the

dendrogram, we identify four classes. We interpret these classes looking at their average position on

the four dimensions found in the PCF (Figure 7). The first class regroups the ‘blurring-by-muting’

RRPs, which are characterized by programmes with low economic salience, in line with Rovny

(2013)’s findings. RRPs in the second class, instead, are characterized by the high salience level of

free-market elements domestically and are protectionist when it comes to international trade.

Therefore, we label them as ‘neoliberal and protectionist’. The third class, ‘pro-welfare’ RRPs,

regroups those parties with programmes including many pro-welfare stances, while other elements

are not particularly stressed. Finally, our fourth class, ‘blurring-by-multiplying’ identifies those

RRPs that stress all economic items, in line with Mudde’s (2007) ‘schizophrenic agendas’ and the

strategy of ‘blurring-by-multiplying’ discussed in Section 2.

To sum up, our results show that differences between RRPs' economic programmes are best

described along four dimensions: ‘welfare’, ‘economic liberalism’, ‘economic management’ and

‘protectionism’. Moreover, we can classify RRPs manifestos in four classes depending on their

economic content: ‘blurring-by-muting’, ‘neoliberal and protectionist’, ‘pro-welfare’, and ‘blurring-

by-multiplying’. These categories will allow us to analyse the evolution of RRPs economic

22

proposals as we compare the positions of different party’s manifestos at different time. In the

following Section: (i) we compare the evolution of Western and Eastern European RRPs on the four

dimensions identified by the PCF and (ii) we describe the evolution of three different RRPs as they

moved from one class to another.

6.2.2 The evolution of RRPs’ positions over time

In order to test our H3, namely whether RRPs modified the composition of their programmes, we

rely on the results from PCF to describe the differences in RRPs’ economic positions over time. To

do so, we distinguish between Eastern and Western European RRPs and, for each of the previously

identified PCF component, we plot the average RRPs’ position over each five-year period (See

Figure 8 and Figure 9). 5

We observe that, for ‘welfare’, both Western and Eastern European RRPs have moved steadily from

relatively low to higher levels of salience during the 20-year period under consideration. The

increase has been faster in Eastern Europe than in Western Europe and, especially, at the turn of the

millennium (See Table 5). Moreover, the decrease in the standard deviation might be a sign of late-

comer RRPs adopting the position of more successful RRPs: a jump on the bandwagon as analysed

by Williams (2006). Overall, these results support our second hypothesis: RRPs have moved from

pro-market positions in the 1990s to a pro-welfare positions in the 2000s.

** Table 5, Figure 8 AND Figure 9 AROUND HERE ***

In the case of ‘economic liberalism’, Figure 8 shows that West European RRPs have decreased the

references to free market policy (without abandoning them), while their Eastern counterparts moved

towards more pro-market pledges following the financial crisis of 2008. This might be an indication

that Western and Eastern European RRPs hold different views about the culprits for the economic

crisis. Figure 9 shows the relative stability of Western European RRPs concerning ‘economic

5 We opt for a five-year average since countries hold elections at different years. A five-year period is however enough

to have at least one data-point (i.e. a manifesto) for each country. Beforehand, we calculated the average position by

party in each five-year period, meaning that parties from countries holding more than one election are computed just

once per period and we calculate.

23

management’ and ‘protectionism’, while Eastern European RRPs have increased their references to

both these terms. Eastern European RRPs now seem to include all dimensions in their political

programme, in line with Mudde's (2007) description of a ‘schizophrenic’ agenda. These results are

also coherent with our H2, to which they provide further evidence, and with Rovny’s (2013, p.20)

conclusions: ‘RRPs maintain a consciously opaque profile on economic issues’. However, the

strategy used by RRPs to blur their position on economic matters seems to have changed. Indeed,

while Rovny results indicated a ‘blurring-by-muting’ strategy, our study shows evidence that RRPs

have been evolving to a blurring-by-multiplying one.

All of these results, read together, confirm the evolution of the RRPs’ strategy on economic matters:

RRPs have been increasing the salience of economic items in their programmes (H1), without taken

a clearer position on economic matters (H2) and modifying the composition of their manifestos

increasing the overall importance of welfare elements (which we consider as a confirmation of our

H3).

These evolutions are difficult to represent for all RRPs on a single graph, due to the large number of

available parties and manifestos in the database. This is why we propose an illustration of these

evolutions relying on three examples: the French FN, the Danish DF, and the Slovakian SNS. We

have chosen these cases as their position changed through time and as this allows us to illustrate the

four clusters identified in Section 6.2.1.

Figure 10 and Figure 11 show the locations of these three major RRPs in the factorial space.

*** Table 5, Figure 10, AND Figure 11 AROUND HERE ***

The French FN clearly moved towards pro-welfare positions. In the 1990s, the party was on the

‘negative’ side of the first axis, while it clearly emphasized redistributive policies in 2017

Presidential election. In relation to ‘economic liberalism’, the FN is more erratic: rather favourable

to liberalism compared to the rest of the party family in 2012, it gives much less attention than

previously to laissez-faire policy. FN shows no clear shift on the ‘economic management’

dimension or on protectionism. Thus, in 2012, compared to the other RRPs, the FN is a welfare-

24

protectionist party with unclear views on economic liberalism. Looking at the clusters: FN

manifestos during the 1990s are classified as ‘neoliberal and protectionist’, which is in line with

Bastow’s (1997) account of the FN economic programme at the time. However, after 2007, party

manifestos fall into the ‘blurring-by-multiplying’ category: free-market references are still present

but are accompanied by a more pro-welfare stance.

The Danish DF also shows an erratic pattern. On the welfare dimension, DF scores are negative for

most of the time span except the 2011 and 2005 elections. However, welfare is not stressed in DF

manifestos in the 2000s. The interpretation is thus less clear than for the French FN. Concerning

‘economic liberalism’, the DF has experienced a decreasing trend in its salience. Overall,

consistently with Rovny (2013), DF appears in our cluster analysis to be an example of blurring-by-

muting: manifestos proposed by the DF fall under the ‘muting’ cluster, apart from 2011 election

when the party stressed pro-welfare statements.

In the 1990-2012 period, the Slovakian SNS covered most of our policy spectrum, with both

positive and negative scores for each dimension depending on the election. Sometimes changes

occurred in a rather limited amount of time. For example, the party did not give much importance to

any economic policies in 1990. In 1994, however, the party ‘rediscovered’ the economy, became

pro-welfare and protectionist, and favoured free-market. Nowadays, the SNS occupies a central

position for each component apart from ‘economic management’ where its negative score points to

a loss in the importance of state control in the party’s ideology. The Slovakian SNS fall under

various classes: as we saw, in 1990 election the SNS opted for a ‘blurring-by-muting’ strategy.

During much of 1990s and until 2010, the party stressed different economic items, proposing an

incoherent program and following a ‘blurring-by-multiplying’ strategy. Finally, more recent

manifestos can be classified as ‘neoliberal’: similarly, to what we saw for Eastern European RRPs,

the Slovakian SNS reacted to 2008 economic crisis by stressing the importance of neoliberal issues

in its programme.

25

6.3 The determinants of economic salience and issue heterogeneity

*** Tables 6 and 7 around here ***

Finally, we turn our attention to the determinants of RRPs strategy both in terms of economic

salience and position blurring. As discussed in Section 3, at this stage, we focus on contextual

explanations related to the political system as such we propose three models for each of our two

dependent variables. In the first model, we provide an opportunistic explanation (model 1): both the

salience and position blurring depend on the strategy adopted by other political parties. In our

second model, we focus on the determinants of party’s success (model 2) and argue that RRP

electoral scores as well as RRPs’ incumbency might influence these parties’ strategy: we expect

larger RRPs, that are closer to competing for government, as well as incumbent RRPs to pay more

attention to economic policy and to adopt a clearer outlook. Finally, we argue that economic

salience might depend on the actual economic outlook (model 3) of the country: as the country’s

economy worsens, non-incumbent RRPs might have incentives in stressing the salience of their

economic proposals as a mean to attack incumbent mainstream parties. Finally, we test a fourth

combined model including all our independent variables.

Table 6 and Table 7 summarise the results of our analysis. Results are rather straightforward in their

interpretation. Regardless of the dependent variable we are analysing, the only covariate which

always turns out to have a statistically significant relationship is the level of salience/heterogeneity

of the political system. Specifically, the level of economic salience in a RRP manifestos strongly

depends on the general level. Our results and the strength of the coefficient both in the opportunistic

and in the combined model seem to indicate that RRPs mostly adapt to change in economic salience

at the systemic level. Despite only at 10% level, in our mainstreaming model we see that partisan

incumbency is linked to an increase of economic salience. This significant relationship disappears

as soon as we consider the level of economic salience registered at the election, however the sign

26

does not change, and it seems to indicate that incumbent RRPs are, in a way, keener to talk about

economic matters than RRPs contesting the election as opposition forces.

The situation does not change if we look at the index of economic issues heterogeneity: we see that

an increase in the system weighted average heterogeneity of 0.1 is roughly translated in an increase

by RRPs of 0.05. In other words, as other parties become less clear in the position on economic

matter, the strategy of position blurring becomes even less costly. Contrarily to the previous case,

none of the covariates turns out to be significant in any model.

To conclude, our preliminary assessment of the possible determinants of RR strategies in terms of

economic salience and blurring seems to indicate that the general level of salience/heterogeneity

characterising an election is the strongest factor. We consider these results to be in line with the

most recent literature on the relationship between RRPs and economic policy (Mudde 2007, Rovny

2013): RRPs mostly compete on non-economic dimensions and are not interested into taking a

stance on these matters unless they are ‘forced’ by other parties’ strategies.

7 Conclusion

This paper explored the evolution of European RRPs' economic programmes since 1990 and

proposed a taxonomy of RRPs’ according to their strategy on economic matters. The diachronic

cross-country approach we used allowed us to describe, classify, and compare the evolution of a

large number of RRPs at the same time, thus connecting together previous researches which

focused either on differences at a given time (De Lange, 2007) or on the evolution of a single party

(Ivaldi, 2015b). Our results can be summarised in four main findings.

First, we highlighted that, consistently with Rovny (2013) and Mudde (2007), economic issues are

not at the core of RRP ideology and are less prominent in their manifestos than in those of other

political families. We showed that RRPs have paid increasing attention to economic matters since

the early 1990s and in particular after the 2008 crisis. We found also that RRPs’ ‘discovery’ of

27

economic issues is characterized by both an increasing salience of these issues and a larger number

of economic topics addressed.

Second, we turned our attention to RRPs’ strategic blurring in order to ascertain whether the

increase in economic salience was accompanied by a clearer position on economic matters. In order

to analyse RR position blurring, we proposed and index of economic issues heterogeneity. Our

results indicate that, despite the increasing attention reserved to economic topics, RRPs continue to

keep the same level of ambiguity as before. This leads us to argue that RRPs have been modifying

their blurring strategy moving from a blurring-by-muting to a blurring-by-multiplicity appeal.

The third series of findings concerns the composition of RRPs’ economic programmes. By the

means of a principal component factor analysis, we were able to illustrate the intra-family economic

policy space along four dimensions. These dimensions describe the salience of ‘welfare’, ‘economic

liberalism’, ‘economic management’ and ‘protectionism’ issues. Our results are in line with the

description of the RR economic policy space already deduced by Otjes et al. (2012) and used by

Ivaldi (2014), despite relying on an inductive approach. Based on these results, we were able to

propose a taxonomy of RRPs as we divided them into four categories depending on the strategy

adopted on economic matters: ‘blurring-by-muting’, ‘neoliberal’, ‘pro-welfare’, and ‘blurring-by-

multiplying’ RRPs. We used the examples of three different parties to identify different strategies

and the evolution of RRPs’ programmes. A blurring-by-muting strategy can be exemplified by the

Danish DF (which was also the last RRP to present a manifesto (in 2001) with no references to

economic issues). Instead, RRPs such as the Slovak SNS adopt what Mudde (2007) calls a

‘schizophrenic agenda’, building manifestos with strong references to both ‘left-wing’ and ‘right-

wing’ economic policies and changing them over time. The French FN has followed a path similar

to that analysed in Ivaldi (2013, 2014, 2015b) moving towards a pro-welfare position and lowering

its initial support for economic liberalism: as of now, the French FN falls under the ‘blurring-by-

multiplying’ class as it combines elements from the different dimensions described in this article.

This pattern is common to most RRPs and there seems to be a converging trend within the party

28

family, in particular if we look at the increasing importance of welfare in RRPs’ manifestos. In

other words, although they might still disagree (for either strategic or ideological reasons) on the

balance between State and Market, the differences between these forces in terms of salience of pro-

welfare issues is decreasing.

Finally, we turned our attention to those factors that could influence the level of economic salience

and position blurring, providing a first account based on three main explanations highlighted by the

literature, namely (i) opportunistic behaviours, (ii) mainstreaming, and (iii) contextual economic

factors. Our preliminary results seem to indicate that RR strategies are first and foremost

determined by those of other political forces. As other parties pay an increasing interest to economic

policy, so do RRPs, and when other political forces increase the vagueness of their stance on

economic policy, RRPs have a clear incentive in blurring even more their position on these topics.

Our study has however some limitations and many dimensions would need further examination.

First, our results need to be combined with results from other sources (Dinas and Gemenis, 2010).

Manifestos are not the only source of policy proposals: parliamentary activity and voting records

(such as Hix, Noury and Roland, 2007), and media analyses (Ellinas, 2010), could be used to

confirm and deepen our analysis. Second, this chapter did not address the determinants of RRPs'

policy choices. Third, our results concern the ‘supply side’ but do not tackle voters’ perceptions of

RR economic programmes. Since RRPs have been undergoing a process of proletarization

(Arzheimer, 2012), the relation between the change in RRPs' electorate' and RRPs' policy proposals

would be a fruitful direction for future research.

29

8 Tables and figures

Item Mean Std.Dev. CV6 Zero

Counts7

401 - Free Market 0.22 0.20 0.905405 43 402 - Incentives 0.29 0.21 0.748252 32 403 - Market Regulation 0.24 0.22 0.932773 45 404 - Economic Planning 0.06 0.13 2.096774 91 405 - Corporatism Mixed Economy 0.02 0.05 2.525114 105 406 - Protectionism : positive 0.13 0.18 1.483051 73 407 - Protectionism : negative 0.04 0.09 2.462604 101 408 - Economic goals 0.22 0.2 0.913242 44 409 - Keynesian Demand Management 0.05 0.11 2.354312 97 410 - Economic Growth : positive 0.18 0.19 1.075145 50 411 - Technology and Infrastructure : positive 0.35 0.23 0.648725 25 412 - Controlled Economy 0.10 0.15 1.525941 76 413 - Nationalisation 0.07 0.13 2.032 88 414 - Economic Orthodoxy 0.20 0.20 0.971014 48 415 - Marxist Analysis 0.01 0.07 7.858672 119 416 - Anti-growth : positive 0.08 0.13 1.703226 87 502 - Culture : positive 0.26 0.22 0.835249 40 503 - Culture : negative 0.29 0.21 0.662069 24 504 - Welfare State Expansion 0.42 0.21 0.525799 16 505 - Welfare State Limitations 0.15 0.18 1.251748 62 506 - Education : positive 0.35 0.21 0.611594 25 507 - Education : negative 0.04 0.01 2.306954 100 701 - Labour groups : positive 0.24 0.21 0.933628 43 702 - Labour groups : negative 0.03 0.08 2.657895 107 Table 1 Summary statistics of MARPOR items covering economic issues (domains 4 - 5 – 7) for RRPs after Prosser’s transformation.

6 Coefficient of variation, defined as the ratio of the standard deviation to the mean. 7 Number of party’s programmes with no references to the item.

30

Party Family8 Obs Mean Std.Dev. CV Min Max

Radical Right 122 0.68 0.19 0.294205 0 0.873

Agrarian 59 0.70 0.18 0.224234 0 0.878

Christian democrat 170 0.76 0.09 0.115625 0.403 0.934

Communist 132 0.75 0.12 0.136187 0 0.905

Conservative 125 0.77 0.08 0.098207 0.239 0.951

Ecologist 85 0.70 0.14 0.206847 0 0.866

Ethnic-regional 85 0.67 0.13 0.193314 0.198 0.856

Liberal 194 0.76 0.10 0.136842 0.167 0.898

Special issue9 53 0.72 0.13 0.209239 0 0.898

Social democratic 204 0.78 0.11 0.135204 0 0.915 Table 2 Economic issues salience by party family 1990-2017 obtained after Prosser’s transformation (2014).

8 The categories are those provided by the MARPOR dataset. 9 The ‘special issue’ category identifies parties with no clear ideological stance and interested by one issue only (e.g. the

Dutch Party for the Animals).

31

Figure 1 Average economic salience by party family 1990-2017

Figure 2 Average number of Econ. Items by party family 1990-2017

Figure 3 Average economic salience by party family in Western Europe 1990-2017

32

Figure 4 Average economic salience by party family in Eastern Europe 1990-2017

Figure 5 Index of economic issues heterogeneity for RRPs 1990-2017

33

Figure 6 Plot of component loadings, first two components.

34

Component Eigenvalue Difference Proportion Cumulative

Proportion Post

rotation

1 3.915 2.35 0.32 0.32 0.21

2 1.563 0.43 0.13 0.45 0.17

3 1.133 0.10 0.09 0.55 0.16

4 1.031 0.19 0.08 0.63 0.10

5 0.839 0.16 0.069 0.70

6 0.679 0.02 0.056 0.76

7 0.656 0.07 0.054 0.81

8 0.579 0.02 0.048 0.86

9 0.557 0.13 0.046 0.91

10 0.418 0.05 0.034 0.94

11 0.364 0.09 0.030 0.97

12 0.265 . 0.022 1

Table 3 PCF Eigenvalue and Variance explained

35

Item 1 2 3 4

401 - Free Market 0.78 402 - Incentives 0.57 403 - Market Regulation 0.5

404 - Economic Planning

0.74 406 - Protectionism : positive

0.94

412 - Controlled Economy

0.80 413 - Nationalisation

0.60

414 - Economic Orthodoxy 0.69 504 - Welfare State Expansion 0.90

505 - Welfare State Limitations 0.77 506 - Education : positive 0.74

701 - Labour groups : positive 0.5 Table 4 Factor loadings after varimax rotation, blanks represent abs. (loadings) <0.50

36

Figure 7 Position of four clusters identified by Ward's method on the economic policy space.

37

Figure 8 The evolution of RRPs economic proposals on 'welfare' (axis 1) and 'economic liberalism' (axis 2) components.

38

Figure 9 The evolution of RRPs economic proposals on 'economic management' (axis 3) and 'protectionism' (axis 4) components.

39

Figure 10 Dimensions of RRPs Econ Salience: welfare (axis 1) and economic liberalism (axis 2).

40

Figure 11 Dimensions of RRPs Econ Salience: economic management (axis 3) and protectionism (axis 4).

41

Years Obs Mean Std.Dev.

1990-1994 28 -0.64 0.98

1995-1999 23 -0.12 1.05

2000-2004 20 -0.00 1.10

2005-2009 20 0.29 0.84

2010-2017 31 0.48 0.675 Table 5 Welfare dimension (axis 1) - Summary Statistics by period. Source: MARPOR and calculation by the author.

42

Model 1 Model 2 Model 3 Model 4

Coeff SE Coeff SE Coeff SE Coeff SE

Average economic salience at election t 1.172** -0.459

1.804*** -0.446

Electoral score at election t

-0.004 -0.003

0.002 -0.003

RRPs is incumbent

0.095** -0.044

0.037 -0.041

GDP Growth

0.004 -0.006 -0.002 -0.006

Constant -0.217 -0.394 0.823*** -0.051 0.770*** -0.045 -0.771** -0.381

R2 0.371

0.266

0.241

0.418 Rho -0.001

0

0.002

-0.002

# of observations 122 122 115 115

Table 6 Explaining economic issue emphasis by RRPs, 1990–2017. Note : results from Prais-Winsten regression correcting for

panel-specific autocorrelation in error terms with panel-corrected standard errors in parentheses. The dependent variable is the

emphasis of economic issues for European RRPs between 1990 and 2017. *** p ≤ 0.01; ** p ≤ 0.05; * p ≤ 0.10.

Model 1 Model 2 Model 3 Model 4

Coeff SE Coeff SE Coeff SE Coeff SE

Party system average economic issues heterogeneity at election t 0.595** -0.251

0.602** -0.258

RRPs is incumbent

0.066 -0.065

0.079 -0.066

Electoral score at election t

0.001 -0.003

0.002 -0.004

GDP Growth

-0.006 -0.007 -0.005 -0.007

Constant 0.274* -0.164 0.646*** -0.055 0.679*** -0.06 0.269 -0.183

R2 0.305

0.265

0.269

0.33 Rho 0.009

0.008

0.009

0.009

# of observations 117 117 111 111

Table 7 Explaining economic issue emphasis by RRPs, 1990–2017. Note : results from Prais-Winsten regression correcting for

panel-specific autocorrelation in error terms with panel-corrected standard errors in parentheses. The dependent variable is the

index of economic issues heterogeneity for European RRPs between 1990 and 2017. *** p ≤ 0.01; ** p ≤ 0.05; * p ≤ 0.10.

43

Economic

Management Welfare Protectionism Economic

Liberalism

MARPOR - planeco 0.60* 0.08 0.12 -0.023

MARPOR - markeco -0.27* 0.01 -0.116 0.62*

MARPOR - welfare -0.01 0.60* -0.09 -0.21

Benoit and Laver (2007) -0.28* -0.43* -0.09 0.66*

Gabel and Huber (2000)- 0.55* 0.15 0.19 -0.64*

Elff (2013) -0.55* 0.08 0.01 0.41*

CHES – L/R economic scale

-0.43* 0.11 -0.21 0.42*

CHES - Salience spending 0.18 0.09 -0.08 -0.25

CHES - Salience deregulation

0.19 0.02 0.05 -0.02

CHES - Salience redistribution

0.23 -0.19 0.02 -0.54*

Table 8 Correlations between the four first components resulting from the PCF and other measures - *= sign. 0.01

44

Figure 12 Evolution of FN ‘Economic Egalitarianism’ salience for Ivaldi (2015, p. 359) and FN ‘Welfare dimension’ obtained in this

paper.

45

Figure 13 Evolution of FN ‘economic liberalism’ salience for Ivaldi (2015, p. 359) and FN ‘economic liberalism’ dimension

obtained in this paper.

46

Figure 14 Evolution of FN ‘economic nationalism’ salience for Ivaldi (2015, p. 159) and FN ’protectionism’ dimension obtained in

this paper.

47

Figure 15 Evolution of FN ‘economic management’ position for Ivaldi (2015, p. 159) and FN ‘economic management’ dimension

obtained in this paper.

48

VAR Factor 1 Factor 2 Factor 3 Factor 4

Welfare

0.7959 Economic Liberalism 0.8767

Economic Management

0.8586 Protectionism

0.653

Untransformed1 0.8848 Untransformed2

0.8322

Untransformed3

0.8068 Untransformed4

0.4642

Untransformed5

0.3678 Table 9 Varimax with Kaiser normalization loadings of second--step factor analysis between dimensions obtained in this paper and

dimensions obtained from untransformed data.

49

VAR Factor 1 Factor 2 Factor 3 Factor 4

Welfare 0.9548 Economic Liberalism

0.9827

Economic Management

0.9826 Protectionism

0.8583

All1 0.8503 All2 0.5636 0.5061

All3

0.7394

All4

0.878 All5

0.5446 0.4377

All6

0.5749 All7

0.4836 Table 10 Varimax with Kaiser normalization loadings of second--step factor analysis between dimensions obtained in this paper and

dimensions obtained from all transformed issues in domain 4 and 5.

50

9 Appendix: list of RRPs

Party Country years N

AN ITA 1994 1

ATAKA BUL 2005, 2009 ,2013, 2014 4

AFD GER 2017 1

BZO AUT 2006, 2008 2

CD NLD 1994 1

DF DEN 1998, 2001, 2005, 2007, 2011 5

EKRE EST 2015 1

ERSP EST 1992 1

Electoral EST 1992 1

FDI-CDN ITA 2013 1

FN FRA 1993, 1997, 2002, 2007, 2012, 2017 6

FPO AUT 1990, 1994, 1995, 1999, 2002, 2006, 2008, 2013 8

FPS SUI 1995 1

HDZ HRV 1990, 1992, 1995, 2000, 2003 5

HSP HRV 1992, 1995, 2000, 2003 4

Jobbik HUN 2010, 2014 2

KPN POL 1991, 1993 2

LAOS GRE 2007, 2012 2

LN ITA 1992, 1994, 1996, 2001, 2006, 2008, 2013 7

LPF NLD 2002, 2003 2

LPR POL 2001, 2005 2

LdT SUI 2007 1

MIP HUN 1998 1

MSI-DN ITA 1992 1

NA LAT 2010, 2011, 2014 3

OP BUL 2017 1

PRM ROM 1992, 1996, 2000, 2004 4

PS FIN 1999, 2003, 2007, 2011 4

PUNR ROM 1990, 1992, 1996 3

PVV NLD 2006, 2010, 2012 3

PX POL 1991 1

SD SWE 2010, 2014 2

SD/DS SUI 1991, 1995, 1999, 2003 4

SNS Slovakia SVK 1990, 1992, 1994, 1998, 2006, 2010, 2012 7

SNS Slovenia SLV 1992, 1996, 2000, 2004, 2008, 2011 6

SPR-RSC CZE 1992, 1996, 1998, 2002 4

SVP/UDC SUI 1991, 1995, 1999, 2003, 2007, 2011 6

TB LAT 1993, 1995 2

UKIP UK 2015 1

VB BEL 1991, 1995, 1999, 2003, 2007, 2010 6

XA GRE 2012 (avg of 2 elections), 2015 (avg of 2 elections) 2

OK CZE 2013 1

Total

122

51

10 Appendix: Robustness and validity

Validity in content analysis refers to the process of verifying that the proposed dimensions measure

what they claim to measure (Bollen, 1989). There are four main types of validity (Ray, 2007): (1)

Content validity: whether the measure we obtain is correct, with reference to the domain we wish to

represent; (2) Convergent/discriminant validity: ‘involves the comparison of alternative measures of

the same concept, and the comparison of measures of different concepts’ (Ray, 2007, p.12); (3)

Criterion or face validity: whether our dimensions are related to the results obtained by other

means; (4) Construct validity: refers to the possibility of obtaining similar results when changing

some of the underlying assumptions. As far as content validity is ensured by our focus on measures

of economic policy (see Sections 1 to 4), next sections tackle the other types of validity criteria.

10.1 Convergent validity

In this section, we measure the extent to which the main components derived from our PCF are

providing – or not – similar results about RRPs' positions on economic matters to those derived

from other types of measures available in the literature. In our case, there are no direct measures of

how RRPs differ from each other in their positions on economic matters. Therefore, we need to

discuss the extent to which the measures proposed in the literature provide results that are consistent

with ours with PCF. Concretely, our results are compared to various deductive approaches:

• The main deductive pre-coded scales for economic issues included in the MARPOR

database (‘planeco’ for planned economy, ‘markeco’ for market economy and ‘welfare’, see

Lave and Budge 1992);

• Benoît and Laver (2006)’s L/R scale for economic issues;

• Non-manifesto related analyses, the most common alternative in the literature being the

Chapel Hill Expert Survey (Bakker et al., 2015).

52

Table 8 presents the correlations between the first four components of our PCF and the measures

obtained using alternative methods.

*** Table 8 AROUND HERE ***

The three pre-coded scales included in the MARPOR database are significantly and positively

correlated to our ‘economic management’, ‘economic liberalism’, and ‘welfare’ scales respectively.

As expected, we also find a that our ‘welfare’ and ‘economic management’ dimensions are

negatively correlated to Benoit and Laver’s (2006) Left/Right scale, while the correlation is positive

with our ‘economic liberalism’ dimension. Finally, we find that our dimensions are mostly

unrelated to expert survey (CHES) formulations of salience, but this might be explained by

discrepancies in coding parties’ positions with respect to MARPOR and partly due to the lack of

data. However, when looking at the only measure available for most of the dataset, the left-right

economic scale, is significantly (although weakly) and consistently correlated to two of our

dimensions: the correlation is positive for ‘economic liberalism’ and negative for ‘economic

management’. Overall, we can thus conclude that our results are robust with respect to ‘convergent

validity’ criterion, bearing in mind the different nature of our (salience) dimensions and the

(confrontational) scales used for comparison: this robustness appears clearly for those dimensions

that clearly identify a left and right economic polarization (‘economic management’ and ‘economic

liberalism’), a little less clearly for the ‘welfare’ dimension and disappears for ‘protectionism’.

10.2 Face validity

*** FIGURES 8 TO 11 AROUND HERE **

In order to consider face-validity, we compare our results for the French FN with those provided by

Ivaldi (2015b). Figure 12 reports the evolution of the FN's position on the ‘welfare’ dimension,

resulting from our factor analysis, and the salience of ‘economic egalitarianism’ in Ivaldi (2015).

The overall trend is similar: FN has moved away from its 1986 – 2012 neo-liberal position. The

comparison is also consistent for the ‘economic liberalism’ dimensions (Figure 13).On the

53

‘economic nationalism and protectionism’ dimension (Figure 14), instead, we obtain opposite

results. However, Ivaldi (2015) includes Eurosceptic references under ‘economic nationalism’,

while we decided not to include the (only) item on the European Union. Finally, on the ‘economic

management’ dimension, Ivaldi does not provide a ‘salience measure’ but rather a positional scale

(Figure 15) whose evolution is not in line with our results. According to our analysis, the FN did

not experience profound changes in this dimension, while Ivaldi finds an important shift towards

‘the left’ in 2012. This might be explained because of the different nature of the two scales (i.e.

salience vs. confrontational).

Overall, the policy space identified by our analysis provides mostly similar trends than that

identified by Ivaldi for the FN, which contributes to validate our approach, with some discrepancies

offering perspectives for further research.

10.3 Construct validity

Finally, we test the extent to which our results depend on our assumptions by changing some of

them. We focus on two choices: the standardised log transformation and the use of Benoit and

Laver (2007) variables subset. We rerun our analyses with two different datasets: in the first, we use

the untransformed variables. In the second we select all (transformed) items in domains 4 and 5

(with the exception of item 501 – ‘Environment’), instead of the variables selected by Benoit and

Laver (2007).

*** Table 9 AND Table 10 AROUND HERE ***

In order to validate our work, we proceed as suggested by Ray (2013) and conduct factor analysis

on parties’ factor scores on our four dimensions and those obtained by the two possible variants.

Table 9 and Table 10 report the rotated results of the analyses. For the analysis with log-

transformation of the data, the main impact appears to be on the hierarchy of the four dimensions,

but three of them (‘welfare’, ‘economic liberalism’ and ‘economic management’) still load strongly

respectively on one and only one factor proving that parties distribution would have been similar.

54

The ‘protectionism’ dimension loads to a lower level. Thus, we can conclude that, while ‘welfare’,

‘economic liberalism’ and ‘economic management’ are all inherent to our dataset, the emergence of

‘protectionism’ derives from the transformation.

Selecting all the variables, leads unsurprisingly to a lower degree of variance explained by each

component. However, again, 3 out 4 of the main components (‘welfare’, ‘economic liberalism’ and

‘protectionism’) in our original results are very correlated to 3 out of 4 of the main components of

this new analysis. In this analysis, the ‘economic management’ component appears less clearly as

an explanatory variable, even though it correlated with the next axis.

Overall, we can conclude that our results are robust with respect to construct validity criterion.

55

11 Bibliography

Aitchison, J., 1986. The statistical analysis of compositional data.

Akkerman, T., 2005. Anti-immigration parties and the defence of liberal values: The exceptional

case of the List Pim Fortuyn. Journal of Political Ideologies, 10(3), pp.337–354.

Akkerman, T., 2012. Comparing Radical Right Parties in Government. Immigration and Integration

Policies in Nine Countries (1996-2010). West European Politics, 35(3), pp.511–529.

Akkerman, T., Lange, S.L. de and Rooduijn, M., 2016. Radical Right-Wing Populist Parties in

Western Europe: Into the Mainstream? Routledge.

Alonso, S. and Volkens, A., 2012. Content-analyzing Political Texts. A Quantitative Approach.

CIS.

Arter, D., 2010. The Breakthrough of Another West European Populist Radical Right Party? The

Case of the True Finns. Government and Opposition, 45(04), pp.484–504.

Arter, D., 2012. Analysing ‘Successor Parties’’: The Case of the True Finns’. West European

Politics, 35(4), pp.803–825.

Arzheimer, K., 2009. Contextual Factors and the Extreme Right Vote in Western Europe, 1980–

2002. American Journal of Political Science, 53(2), pp.259–275.

Arzheimer, K., 2012. Working Class Parties 2.0? Competition between Centre Left and Extreme

Right Parties. In: J. Rydren, ed., Class Politics and the Radical Right. London: Routledge, pp.75–

90.

Arzheimer, K. and Carter, E., 2006. Political Opportunity Structures and Right-Wing Extremist

Party Success. European Journal of Political Research, 45, pp.419–443.

Bakker, R., Edwards, E., Hooghe, L., Jolly, S., Koedam, J., Kostelka, F., Marks, G., Polk, J.,

Rovny, J., Schumacher, G., Steenbergen, M., Vachudova, M.A. and Zilovic, M., 2015. 1999-2014

Chapel Hill Expert Survey Trend File. Chapel Hill: Chapel Hill, NC: University of North Carolina.

Bastow, S., 1997. Front national economic policy: From neo‐liberalism to protectionism. Modern &

Contemporary France, 5(1), pp.61–72.

Benoit, K. and Laver, M., 2006. Party policy in modern democracies. Routledge.

Benoit, K. and Laver, M., 2007. Estimating party policy positions: Comparing expert surveys and

hand-coded content analysis. Electoral Studies, 26(1), pp.90–107.

Betz, H.-G., 1994. Radical Right-Wing Populism in Western Europe. Houndmills, London:

Macmillan.

56

Beyme, K., 1988. Right-Wing Extremism in Post-War Europe. In: K. Beyme, ed., Right-Wing

Extremism in Western Europe. London: Frank Cass, pp.1–18.

Bollen, K., 1989. Structural Equations with Latent Variables John Wiley New York.

Bornschier, S. and Kriesi, H., 2012. The populist right, the working class, and the changing face of

class politics. Class politics and the radical right, pp.10–30.

Braun, D., Mikhaylov, S., Schmitt, H., Popa, S.A. and Dwinger, F., 2016. European Parliament

Election Study 2014, Manifesto Study. GESIS Date Archive, Cologne. ZA5057 Data file Version

1.0. 0, doi: 10.4232, 1.

Budge, I., 1999. Estimating Party Policy Positions: From Ad Hoc Measures to Theoretically

Validated Standards. Joint Sessions of the European Consortium for Political Research, Workshop.

.

Budge, I., 2001. Mapping policy preferences: estimates for parties, electors, and governments,

1945-1998. Oxford University Press on Demand.

Castles, F.G. and Mair, P., 1984. Left-Right Political Scales: Some ‘Expert’ Judgments. European

Journal of Political Research, 12(1), pp.73–88.

Cole, A., 2005. Old right or new right? The ideological positioning of parties of the far right.

European Journal of Political Research, 44(2), pp.203–230.

De Koster, W., Achterberg, P. and Van der Waal, J., 2013. The new right and the welfare state: The

electoral relevance of welfare chauvinism and welfare populism in the Netherlands. International

Political Science Review, 34(1), pp.3–20.

De Lange, S.L., 2007. A new winning formula? The programmatic appeal of the radical right. Party

Politics, 13(4), pp.411–435.

Dinas, E. and Gemenis, K., 2010. Measuring parties’ ideological positions with manifesto data a

critical evaluation of the competing methods. Party politics, 16(4), pp.427–450.

Dolezal, M., Ennser-Jedenastik, L., Müller, W.C. and Winkler, A.K., 2014. How parties compete

for votes: A test of saliency theory. European Journal of Political Research, 53(1), pp.57–76.

Downs, A.S., 1957. An Economic Theory of Democracy. New York: Harper and Row.

Ellinas, A.A., 2010. The media and the far right in Western Europe: Playing the nationalist card.

Cambridge University Press Cambridge.

Franzmann, S. and Kaiser, A., 2006. Locating Political Parties in Policy Space A Reanalysis of

Party Manifesto Data. Party politics, 12(2), pp.163–188.

Gabel, M.J. and Huber, J.D., 2000. Putting parties in their place: Inferring party left-right

ideological positions from party manifestos data. American Journal of Political Science, pp.94–103.

Gemenis, K., 2013. What to Do (and Not to Do) with the Comparative Manifestos Project Data:

Comparative Manifestos Project Data. Political Studies, 61, pp.3–23.

57

Hernández, E. and Kriesi, H., 2016. The electoral consequences of the financial and economic crisis

in Europe. European Journal of Political Research, 55(2), pp.203–224.

Hix, S., Noury, A.G. and Roland, G., 2007. Democratic politics in the European Parliament.

Cambridge: Cambridge University Press.

Ignazi, P., 2003. Extreme right parties in Western Europe. Oxford; New York: Oxford University

Press.

Ivaldi, G., 2013. Vers un nouveau chauvinisme du welfare? La transformation du programme

économique du Front national (1984-2012). Congrès AFSP Paris 2013. .

Ivaldi, G., 2014. A new course for the French radical-right? [online] Available at:

<http://hal.archives-ouvertes.fr/halshs-01059581/>.

Ivaldi, G., 2015a. Du néolibéralisme au social-populisme? In: Les faux-semblants du Front

national. Presses de Sciences Po, pp.161–184.

Ivaldi, G., 2015b. Towards the median economic crisis voter? The new leftist economic agenda of

the Front National in France. French Politics, 13(4), pp.346–369.

Ivarsflaten, E., 2005. The vulnerable populist right parties: No economic realignment fuelling their

electoral success. European Journal of Political Research, 44(3), pp.465–492.

Kaiser, H.F., 1958. The varimax criterion for analytic rotation in factor analysis. Psychometrika,

23(3), pp.187–200.

Keskinen, S., 2016. From welfare nationalism to welfare chauvinism: Economic rhetoric, the

welfare state and changing asylum policies in Finland. Critical Social Policy, 36(3), pp.352–370.

Kim, H. and Fording, R.C., 1998. Voter ideology in Western democracies, 1946–1989. European

Journal of Political Research, 33(1), pp.73–97.

Kitschelt, H., 1995. The Radical Right in Western Europe: A Comparative Analysis. Ann Arbor:

University of Michigan Press.

Kitschelt, H., 2004. Diversification and reconfiguration of party systems in postindustrial

democracies. Bonn: Friedrich Ebert Stiftung.

Kriesi, H., Grande, E., Lachat, R., Dolezal, M., Bornschier, S. and Frey, T., 2008. West European

politics in the age of globalization. Cambridge University Press Cambridge.

Laver, M., 2001. Position and salience in the policies of political actors. Estimating the policy

position of political actors, 20, p.66.

Lowe, W., Benoit, K., Mikhaylov, S. and Laver, M., 2011. Scaling Policy Preferences from Coded

Political Texts: Scaling Policy Preferences from Coded Political Texts. Legislative Studies

Quarterly, 36(1), pp.123–155.

McDonald, M.D. and Mendes, S.M., 2001. Checking the party policy estimates: convergent

validity. In: I. Budge, H.-D. Klingemann, A. Volkens, J. Bara and E. Tanenbaum, eds., Mapping

political preferences. Estimates for parties, electors, and governments, 1945-1998. Oxford: Oxford

University Press, pp.127–142.

58

Minkenberg, M., 2015. Transforming the Transformation?: The East European Radical Right in the

Political Process. Routledge.

Mudde, C., 2007. Populist radical right parties in Europe. Cambridge: Cambridge Univ. Press.

Mudde, C., 2014. The far right and the European elections. Current History, 113(761), p.98.

Norris, P., 2005. Radical Right. Voters and Parties in the Regulated Market. Cambridge, New

York: Cambridge University Press.

Petry, F. and Pennings, P., 2006. Estimating the Policy Positions of Political Parties From

Legislative Election Manifestos 1958–2002. French Politics, 4(1), pp.100–123.

Prosser, C., 2014. Building policy scales from manifesto data: A referential content validity

approach. Electoral Studies, 35, pp.88–101.

Raunio, T., 2012. “Whenever the EU is Involved, You Get Problems”?: Explaining the European

Policy of the (True) Finns. SEI Working Paper 127. Brighton: Sussex European Institute.

Ray, L., 2007. Validity of measured party positions on European integration: Assumptions,

approaches, and a comparison of alternative measures. Electoral Studies, 26(1), pp.11–22.

Rovny, J., 2013. Where do radical right parties stand? Position blurring in multidimensional

competition. European Political Science Review, 5(01), pp.1–26.

Rydgren, J., 2004a. Explaining the Emergence of Extreme Right-wing Populism. The Case of

Denmark. West European Politics, 27, pp.474–502.

Rydgren, J. ed., 2004b. The Populist Challenge. Political Protest and Ethno-Nationalist

Mobilization in France. New York: Berghahn Books.

Rydgren, J., 2007. The sociology of the radical right. Annu. Rev. Sociol., 33, pp.241–262.

Singer, M.M., 2013. The global economic crisis and domestic political agendas. Electoral Studies,

32(3), pp.404–410.

Spoon, J.-J., Hobolt, S.B. and Vries, C.E., 2014. Going green: Explaining issue competition on the

environment. European Journal of Political Research, 53(2), pp.363–380.

Volkens, A., Lehmann, P., Matthieß, T., Merz, N., Regel, S. and Werner, A., 2015. The Manifesto

Data Collection. Manifesto Project (MRG/CMP/MARPOR). Version 2015a. Wissenschaftszentrum

Berlin für Sozialforschung (WZB).

Ward Jr, J.H., 1963. Hierarchical grouping to optimize an objective function. Journal of the

American statistical association, 58(301), pp.236–244.

Williams, M.H., 2006. The Impact of Radical Right-Wing Parties in West European Democracies.

Palgrave Macmillan.

Wold, S., Esbensen, K. and Geladi, P., 1987. Principal component analysis. Chemometrics and

Intelligent Laboratory Systems, 2(1–3), pp.37–52.